A universal graph deep learning interatomic potential for the periodic table

By Chi Chen, Shyue Ping Ong. Nature Computational Science, 2022

1,004 citations recorded by the source corpus.

Interatomic potentials (IAPs), which describe the potential energy surface of atoms, are a fundamental input for atomistic simulations. However, existing IAPs are either fitted to narrow chemistries or too inaccurate for general applications. Here we report a universal IAP for materials based on graph neural networks with three-body interactions (M3GNet). The M3GNet IAP was trained on the massive database of structural relaxations performed by the Materials Project over the past ten years and has broad applications in structural relaxation, dynamic simulations and property prediction of materials across diverse chemical spaces. About 1.8 million materials from a screening of 31 million hypothetical crystal structures were identified to be potentially stable against existing Materials Project crystals based on M3GNet energies. Of the top 2,000 materials with the lowest energies above the convex hull, 1,578 were verified to be stable using density functional theory calculations. These results demonstrate a machine learning-accelerated pathway to the discovery of synthesizable materials with exceptional properties. A machine learning interatomic potential model is designed and trained on diverse crystal data comprising 89 elements, enabling materials discovery across a vast chemical space without retraining.

Subfields: Computational Theory and Mathematics, Materials Chemistry

View Paper: https://www.semanticscholar.org/paper/450bb2530e4fb48e828161d3e7741c5b4081c9a2

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Materials Chemistry

Citations in the corpus

A foundation model for atomistic materials chemistry. (2023, neutral): “…using equivariant graph neural network architectures with the capacity to compute forces were also trained on MP-based datasets, including M3GNet ( 23 ) and CHGNet ( 24 ), which were both trained on snapshots from DFT relaxations of the MP structures, with the CHGNet using the MPtrj dataset…”
A generative model for inorganic materials design (2025, positive): “In recent years, the combination of generating structures by RSS, substi-tution or evolutionary methods with MLFFs has proven successful in exploring chemical systems 21,23,43 .”
Scaling deep learning for materials discovery (2023, methodology): “62, a NequIP model was trained on the M3GNet dataset.”
Crystal Diffusion Variational Autoencoder for Periodic Material Generation (2021, methodology): “Another related method is to train a force ﬁeld from QM forces and then apply the learned force ﬁeld to generate stable materials by minimizing energy (Deringer et al., 2018; Chen & Ong, 2022).”